Does everyone believe FSD will be viable in the next 3 years? 5 years? 10 years?

[showmemo]

I love the idea of FSD for some scenarios including trips. I also have a son that may never drive, so having full FSD would provide freedom to both of us... but, without other manufacturers pushing as hard as Tesla, I wonder if/when it will truly be a fully working product.

I understand that Tesla has big plans for FSD with Robotaxi's and such, but I would think that they will be delayed until they have more companies pushing along side them to open up regulations.

I ordered it with my Tri-Motor since it will be 2 yrs or so until I get it... if I was ordering one for delivery in the next month or 2, I probably wouldn't pay the premium.

What say you?

Advertisement

 

[Crissa]

I would get it, to have the option and expanded Autopilot features until then.

I'm a big proponent of future-proofing my purchases; I tend to use things like laptops for five years, so I look at what my needs in five years will be. And then I skimp on things which are readily available (like disk space)

Price becomes a tertiary concern unless it's confining. (Which is has been in the past) I would rather fit future features and support rather than things I can replicate myself (like color and nice seats or even performance mods)

-Crissa

 

[Ehninger1212]

Im not sure where it will be, but I know I am planning on keeping my CT.. indefinitely.. I'll probably be buried in it.. me and my CT 6' under for eternity... Having said that I will Probably by the FSD package just for "Future proofing" as @Crissa Said.

 

[lancethibault]

Just thinking/typing out loud here.
I think what the vehicle is capable of vs what the laws allow, both go into the discussion of defining FSD. To me, FSD means the vehicle and laws support the vehicle doing it all. e.g. I can sleep in the car traveling across the country and I do not have to worry about stopping to pay a toll or going through construction zones, nor will I be pulled over for doing so. I don't see this becoming reality in the next 20 years.

To your point about allowing someone who cannot drive, to essentially be taxied by the car, I'm not sure I personally recognize a difference. If a person is not physically capable to control the car they are "driving", then from a law perspective, they might as well be sleeping.

Having said that, once the vehicle manufactures begin touting their FSD capability, I could see laws for exception put in place for specific parts of the population, (e.g. the blind). Although I can't image this without some kind of high cost insurance until there are enough vehicles conducting FSD for long enough time that the actuaries can determine one way or the other that FSD is either less or more risk to the insurance company...and by default the population. I could see this type of FSD by exception (where the driver does not legally have to maintain vehicle control in some way) possibly in the next ten years.

Yes, my pre-order has FSD and based on a late 21' production start, I would expect delivery mid 22' and my expectation is that I will not be able to use FSD as I describe it due mostly to the laws and only partially due to the vehicle capability. I think the 90% solution for the capability will be there by then and about 10% legally (I'm not even sure what the current laws are for FSD, but it's my assumption that it's not legal...except for maybe the hands free parallel parking that many cars seem capable of now.) The last 10% of vehicle capability will also be difficult to achieve. I'm not talking about off-road driving where there are no roads but only trails and boulders between the trails. For that I'm not sure FSD will be a thing in my lifetime, and why would we want it to be? But that last 10% I am talking about is for areas that don't have much for signs. Where the county roads are dirt and not always labeled well, and sometimes they are seasonal...or when driveways are also kind of like long county dirt roads.

Laws, insurance, and that last 10% of FSD capability I don't see coming together before 2030. I would love to be wrong.

 

[Sirfun]

I love the idea of FSD for some scenarios including trips. I also have a son that may never drive, so having full FSD would provide freedom to both of us... but, without other manufacturers pushing as hard as Tesla, I wonder if/when it will truly be a fully working product.

I understand that Tesla has big plans for FSD with Robotaxi's and such, but I would think that they will be delayed until they have more companies pushing along side them to open up regulations.

I ordered it with my Tri-Motor since it will be 2 yrs or so until I get it... if I was ordering one for delivery in the next month or 2, I probably wouldn't pay the premium.

What say you?

I'm in a similar situation to you. I have a 23 yr old son who is a high functioning special needs person. If we had dependable FSD he could pursue a drivers license and drive. Without it, a drivers license and driving in Southern California is not a good idea. So I am also anxiously watching the development of FSD. I find it interesting so many people have already purchased FSD and are committed to it being viable. Generally speaking I try not to buy things I'm not going to get good use of. I'm buying the CT thinking it's going to be something I enjoy driving and I have no desire to hand over the wheel to FSD. So if I were configuring my CT for delivery today I wouldn't include FSD. However, this weekend I added my second reservation for a Tri-motor because of the added range and when I saw my reservation number was over 1,200,000 I included FSD because I am hoping FSD will be reliable by the time that CT is ready to be delivered.
So Scott, for both our benefit I really hope dependible FSD does become a reality soon! My first reservation of a Dual motor doesn't have FSD included, but when it comes time to configure it for delivery if FSD has improved considerably I may purchase it. That's the beauty of Tesla building it into all their vehicles, it's always an easy option. One other thing I want, is the ability to lock the CT into a mode like valet mode, where the acceleration would be very limited.

 

[Balthezor]

Just in my opinion, based on regulations, car insurance, ICE car manufacturers fighting against it, one accident with robotaxi will set it back years, etc. I think we are 10 years out. I cannot see it happening in less than 5 years like most people are hoping with robotaxis.

I could be wrong. It be great if I am wrong on this one. As a Model Y owner.

 

[Crissa]

Just in my opinion, based on regulations, car insurance, ICE car manufacturers fighting against it, one accident with robotaxi will set it back years, etc. I think we are 10 years out. I cannot see it happening in less than 5 years like most people are hoping with robotaxis.

It does seem that expectations and the requirements are unreasonably high. (And the comparisons with human drivers very lenient to the human drivers).

-Crissa

 

[Hoppi]

And the comparisons with human drivers very lenient to the human drivers.

Ummm..... This looks like a cryptic way of saying that most humans are horrible drivers. LOL. I don't actually disagree.

Seems the requirement that FSD be better than human drivers isn't quite as hard as most people think. If we compare by statistics FSD will be better than human drivers much earlier than most people will feel comfortable using it. Even if statistically FSD is better, there will most certainly be the occasional accident while using FSD. Of course most people will not do a logical rational comparison with all the human caused accidents though. So it will take much longer for general acceptance to kick in I think. And government can be notoriously slow about some things so who knows how long it will take government to permit FSD use by civilians.

Now once FSD does get permitted, I wonder how long it will take for government to then mandate FSD if it is truly safer than human control. I imagine the logic will be that in certain specific accidents where FSD was available but not being used, if the end result was serious injury or death of an innocent party, then not using the FSD was essentially vehicular manslaughter.

So for FSD I see 3 big milestones coming at some point:

1. FSD declared to be developed and working successfully.
2. FSD permitted to be used by the general public.
3. FSD mandated for use in all available cases. (Clearly vehicles without FSD would be grandfathered in as not requiring FSD).

Once FSD becomes mandated our roads will probably be MUCH safer than they are now. But there will be a huge fight coming over step 3 as there will be quite a number of people who will strongly resist it.

 

[Crissa]

I assume step three will mean more and more areas would be off-limits to human drivers.

And that might come really quick once we get past step 2... as stats start backing up claims like Elon makes about safety. It doesn't take much to beat human drivers on average.

-Crissa

 

[Hoppi]

I assume step three will mean more and more areas would be off-limits to human drivers.

And that might come really quick once we get past step 2... as stats start backing up claims like Elon makes about safety. It doesn't take much to beat human drivers on average.

-Crissa

Yes, it will probably be rolled out fairly quickly. Step 2 will take public acceptance as politicians will need to be assured that it won’t be held against them in case of failure. But step 3 will have lots of big money corporate backing (especially from insurance) and will proceed despite individuals protesting.

 

[Dids]

There is another huge driver of FSD. Trucking. It will provide an economic incentive for trucking companies and once they are doing it who is going to say no to passenger cars?
So I predict the amount of time between step 1 and step 2 to be very short.
Additionally the politically powerful old people will be rapid adopters for the first time in history.

 

[TheLastStarfighter]

I think highway FSD is pretty much there now. Within 3 years I think it will work extremely well. City driving is a whole other game. Far, far more complicated. When I'm driving downtown, my head is on a swivel. When I'm on the highway, I can practically fall asleep. I don't know how you measure that in computations, but I would guess the processing and programing requirements are at least a factor of 10. So I think we're about 10 years from reliable, consistent urban driving. Of course, recent tech like cel phones and the internet advanced way faster than I expected, so it's possible we could be handsfree downtown in about 5 years.

 

[lqdchkn]

IMO within 3 years for sure.

 

[ajdelange]

I think highway FSD is pretty much there now.

Which manufacturer has that? It certainly isn't Tesla. I'm guessing you have never "driven" a Tesla with autopilot or you wouldn't have made this comment.

What people don't seem to understand is that autopilot is a machine. It is very, very good at tirelessly doing the mindless repetitive tasks that driving requires. Just maintaining speed, separation from the vehicle in front of you and staying in lane requires hundreds of control inputs from the driver over even a mile. The machine can do that. Allowing it to do so reduces driver fatigue greatly and improves safety because the machine isn't subject to fatigue. The problem comes in when the machine is required to exercise judgement. It hasn't any. It cannot reason - only make Bayseian decisions based in what it observes through its relatively crude "eyes" and on the a-prioris that have been trained into it. Thus while at the end of a trip the driver who used autopilot will be more rested than if he didn't he will probably have had to wrest control from it at least once per hour.

Within 3 years I think it will work extremely well.

It already works extremely well if you limit your expectations to what is reasonable to expect. If you expect it to replace a human driver it works very poorly and will continue to do so for a very long time.

 

[TheLastStarfighter]

Which manufacturer has that? It certainly isn't Tesla. I'm guessing you have never "driven" a Tesla with autopilot or you wouldn't have made this comment.

What people don't seem to understand is that autopilot is a machine. It is very, very good at tirelessly doing the mindless repetitive tasks that driving requires. Just maintaining speed, separation from the vehicle in front of you and staying in lane requires hundreds of control inputs from the driver over even a mile. The machine can do that. Allowing it to do so reduces driver fatigue greatly and improves safety because the machine isn't subject to fatigue. The problem comes in when the machine is required to exercise judgement. It hasn't any. It cannot reason - only make Bayseian decisions based in what it observes through its relatively crude "eyes" and on the a-prioris that have been trained into it. Thus while at the end of a trip the driver who used autopilot will be more rested than if he didn't he will probably have had to wrest control from it at least once per hour.

It already works extremely well if you limit your expectations to what is reasonable to expect. If you expect it to replace a human driver it works very poorly and will continue to do so for a very long time.

I would call one input per hour to be "pretty much there". At the rate they are progressing, 3 years seems like a reasonable time to have it working "extremely well", and by that I mean a high percentage of typical highway trips will work without intervention. Since the vast majority of highway accidents are from human error, especially losing focus/distraction/operating under influence, I think FSD will be superior to humans, overall, at that time.